Abstract
PURPOSE: Most of the previous meta-analyses examining the effects of non-steroidal anti-inflammatory drugs (NSAIDs) on preventing heterotopic ossification (HO) following total hip arthroplasty (THA) have separately analyzed Brooker's classification I, II, III, and IV which may misrepresent their ordinal nature. We therefore applied a Bayesian network meta-regression that incorporates the ordinal nature of Brooker's classification to more robustly assess NSAID efficacy and determine the optimal regimen. METHODS: We searched the Cochrane Library, Embase, and PubMed for randomized controlled trials (RCTs). Cumulative regressions were conducted for ordinal variants to generate Napierian Logarithm odds ratios (lnOR) and the standard error of lnOR (selnOR) for each study. Subsequently these data were used to conduct Bayesian network meta-analysis and further network meta-regression to generate pairwise ORs, showing pairwise effect sizes (ESs). RESULTS: 17 studies (5436 patients,14 regimens) were eligible. In the raw data analysis, celecoxib 400 mg/d, etoricoxib 90 mg/d, ibuprofen 1200 mg/d, indomethacin 75 mg/d, indomethacin 100 mg/d, indomethacin 150 mg/d, meloxicam 7.5 mg/d, meloxicam 15 mg/d, and naproxen 750 mg/d were conspicuously effective (OR: 0.048 ~ 0.351) compared with placebo. The ESs were comparable among these regimens except for ibuprofen 1200 mg/d, which was inferior to indomethacin 100 mg/d (OR = 0.382, 95%CI: 0.171 to 0.887) and indomethacin 150 mg/d (OR = 0.136, 95%CI: 0.020 to 0.970). In the network meta-regression analysis, after adjusting for follow-up time, the significance of diclofenac 150 mg/d (OR = 0.102, 95%CI: 0.013 to 0.835) emerged compared with placebo. The results of effective regimens aforementioned resembled the initial findings (OR: 0.039 ~ 0.249). All the effective agents, including diclofenac 150 mg/d, were comparable in ESs. CONCLUSIONS: Considering the efficacy of preventing HO following THA observed in our research, together with analgesic effect and gastrointestinal tract safety from previous literature, etoricoxib 90 mg/d is recommended as the optimal choice for patient undergoing THA. More head-to-head and long-term studies are needed.